No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Content-Based Health Recommender System for ICU Patient
Abstract
In this study, the authors propose a generic architecture, associated terminology and a classificatory model for observing ICU patient's health condition with a Content-Based Recommender (CBR) system consisting of K-Nearest Neighbors (KNN) and Association Rule Mining (ARM). The aim of this research is to predict or classify the critically conditioned ICU patients for taking immediate actions to reduce the mortality rate. Predicting the health of the patients with automatic deployment of the models is the key concept of this research. IBM Cloud is used as Platform as a Service (PaaS) to store and maintain the hospital data. The proposed model demonstrates an accuracy of 95.6% from the KNN Basic 'ball tree' algorithm. Also, real-time testing of the deployed model showed an accuracy of 87% while comparing the output with the actual condition of the patient. Combining the IBM Cloud with the Recommender System and early prediction of the health, this proposed research can provide a complete medical decision for the doctors.
... Because kd tree resides in memory after it is trained, it can also be used in service layer to provide querying service. Compared to other kNN(K Nearest Neighbor) related algorithms such as brute kNN and ball tree [19], [20], kd tree has its advantages in performance. Other models such as linear regression, neural network, SVM and so on can also be used in in batch and realtime layers. ...
... When n is large the brute kNN is not a practicable method. Many improved methods are designed to solve this problem such as the kd tree [17], [18], ball tree [19], [20], Hybrid Spill Tree [25], [26] and so on. The time complexity of these tree-based methods are O(log(n)). ...
... The testing dataset includes 1000 samples and all models are trained by the same training dataset. When k=3, 5,10,15,20,25 we compute the averaging prediction time of 1000 samples on brute kNN, ball 6 In this section the time and space performance of brute kNN, kd tree and ball tree are compared. The results show that kd tree is better than other two models in our application. ...
In this paper a system is designed and implemented to predict the price of second-hand housing. This system based on Lambda architecture can execute prediction in both real-time and batch modes so it can give two kinds of different price predictions that reflect current and historical conditions respectively. The kNN related algorithms are used for price prediction. By comparing the performance of brute kNN, kd tree and ball tree, kd tree is selected as the price prediction model of the system. In system implementation the kd tree model is chosen to predict prices in both real-time and batch services. The kd tree model can also recommend housings to user besides price prediction. The experiment shows the effectiveness of our system. Time and space performance of brute kNN, kd tree and ball tree are compared by experiments. And the evaluation metrics of other available maching learning models are compared. The reason of choosing the kd tree model is also explained by the experimental results.
... Cosine Similarity Metric: It measures the similarity between various documents and refers to the cosine of the angle between two vectors (each document is modeled in the vector space) (Neloy, Oshman, Islam, Hossain, & Zahir, 2019). ...
... Learning-based models are supervised models (M. Sharma & Ahuja, 2017) or unsupervised models (Neloy et al., 2019;Folino & Pizzuti, 2010), while mathematical-based models consist Vector space model (Ju & Zhang, 2021;Agarwal & Mostafa, 2011) and statistical analysis-based models (Marcu, 2020;Mustaqeem, Anwar, Khan, & Majid, 2017), are explained in this subsection. M. Sharma and Ahuja (2017) suggested a health care recommendation system that could provide the related information considering various treatments per patient's requirements and limitations. ...
Background
The maintenance of health through prevention, diagnosis, and treatment of illnesses, injuries, and other mental and physical impairments in people is called healthcare that is performed by professionals in health fields. All activities performed in providing primary, secondary, and tertiary care, as well as in public health, are components of healthcare systems that are required to analyze massive amounts of patients’data to gain insights and aid in disease prediction. The healthcare recommender system (HRS) is a growing and robust platform for healthcare services.
Objective
The purpose of this paper is to identify, taxonomically classify, and compare current HRS researches in a systematic way.
Method
This study presents a systematic literature review (SLR) method for HRSs regarding 41 papers published between 2010 and 2021; the selected articles fall into five classes: collaborative-based, content-based, knowledge-based, context-based, and hybrid.
Results
Seventeen essential factors are identified to examine the selected HRSs, and each of the papers were studied to find advantages, disadvantages, evaluation types, and tools. In addition, we address a discussion of challenges, future directions, and open issues.
Conclusion
This paper demonstrates that HRSs are still in their early stages of maturity, but they are developing. It identifies the need for a scalable, confidential, and reliable framework to aid in the improvement of both patients’ and health professionals’ trust in HRSs. This research reveals a lack of standardization of medical codes across various healthcare systems. In addition, this review identifies the need for novel patterns to address cold start problems.
... The suggested technique is adequate for recommending citations without using meta-data and achieves an 18% F1 score in the top 20 recommendations. Similarly, Asif et al. [23] proposed a KNN and Association Rule Mining (ARM) to forecast serious condition ICU patients to lessen mortality rates. For the management of hospital data, the IBM cloud platform is employed. ...
A massive amount of information explosion over the internet has caused a possible difficulty of information overload. To overcome this, Recommender systems are systematic tools that are rapidly being employed in several domains such as movies, travel, E-commerce, and music. In the existing research, several methods have been proposed for single-user modeling, however, the massive rise of social connections potentially increases the significance of group recommender systems (GRS). A GRS is one that jointly recommends a list of items to a collection of individuals based on their interests. Moreover, the single-user model poses several challenges to recommender systems such as data sparsity, cold start, and long tail problems. On the contrary hand, another hotspot for group-based recommendation is the modeling of user preferences and interests based on the groups to which they belong using effective aggregation strategies. To address such issues, a novel “KGR” group recommender system based on user-trust relations is proposed in this study using kernel mapping techniques. In the proposed model, user-trust networks or relations are exploited to generate trust-based groups of users which is one of the important behavioral and social aspects. More precisely, in KGR the group kernels and group residual matrices are exploited as well as seeking a multi-linear mapping between encoded vectors of group-item interactions and probability density function indicating how groups will rate the items. Moreover, to emphasize the relevance of individual preferences of users in a group to which they belong, a hybrid approach is also suggested in which group kernels and individual user kernels are merged as additive and multiplicative models. Furthermore, the proposed KGR is validated on two different trust-based datasets including Film Trust and CiaoDVD. In addition, KGR outperforms with an RMSE value of 0.3306 and 0.3013 on FilmTrust and CiaoDVD datasets which are lower than the 1.8176 and 1.1092 observed with the original KMR.
... In addition, this approach contains attributes matched with the user profile and past preferred history of the user. For this case, the collected documents are filtered to obtain the keywords for the recommendation using content-based methods (Islam et al. 2019). ...
Internet of things enabled wireless sensor networks using fuzzy logic approach is used to collect health metrics of individuals, make real-time analyses, and make efficient therapeutic decisions on Covid-19 cases while remotely monitoring the individuals and reducing the virus's consequences. Symptom data for Covid-19 patients is collected using wearable sensors. Data fusion is performed using a fuzzy logic classifier. Preprocessing and filtering of data to produce a verdict are carried out in data fusion. The data with a possible decision is saved in cloud infrastructure, making it accessible to anyone, including users, medical professionals, and local hospitals. The suggested procedure surpasses other similar methods such as logistic regression, decision tree, naïve bayes, k-nearest neighbor, and support vector machine relating to the accurateness, error rate, F-measure, and ROC area according to experimental results. Complicated decisions in the medical industry could be made more successfully with the support of the suggested technique to predict COVID-19 based on fuzzy logic.
... In addition, this approach contains attributes matched with the user profile and past preferred history of the user. For this case, the collected documents are filtered to obtain the keywords for the recommendation using content-based methods (Islam et al. 2019). ...
Traditional irrigation systems for agricultural lands are expensive, time-consuming, and labor-intensive. Utilizing cutting-edge technology like machine learning, the internet of things, and wireless sensor networks, smart farming addresses current issues with agricultural sustainability while boosting the quantity and quality of crop production from the fields to fulfill the rising food demand. Soil moisture and temperature sensors are used to create a low-cost, real-time IoT-based automatic irrigation system. Two groups have been formed with the sensor information such as “require water” and “not require water” and saved on the server. The device intelligently determines whether the field needs water and automatically turns “ON” or “OFF” the motor. Machine learning based models such as k-nearest neighbor, support vector machines, decision tree, and naive bayes are applied to decide irrigation requirements. Performance metrics show that the KNN classifier performs better than the other two models. The suggested framework allows for better field monitoring and visualization.
... In addition, this approach contains attributes matched with the user profile and past preferred history of the user. For this case, the collected documents are filtered to obtain the keywords for the recommendation using content-based methods (Islam et al. 2019). ...
In wireless sensor networks (WSN), clustering is one of the most significant ways of extending the network’s lifetime. A clustering protocol effectively ensures that nodes save energy and extend the network’s lifespan. Due to cluster creation overhead and cluster head duties, most present clustering techniques require much energy. An energy-efficient adaptive clustering and routing strategy to extend the longevity of WSN is proposed A novel weight-based LEACH (LEACH-W) protocol is proposed to enhance the standard LEACH protocol that saves energy and extends the network’s life. The proposed technique reduces energy usage in WSN by lowering the distance for node-to-node data transmission using a uniform cluster distribution and adaptive multi-hop techniques. The distance between the nodes is determined, and the residual energy of sensor nodes is considered while choosing the cluster head (CH) nodes, and the number of nodes is computed for every CH. The suggested solution addresses variations in network conditions and considers changeable settings in real-world situations. The suggested solution extends the lifespan of the WSN through load-balanced clustering. The impact of changing weights on energy minimization was studied, and the ideal weight energy was estimated. The number of nodes is also determined for each CH in the proposed method. Simulation findings reveal that the suggested clustering algorithm and routing protocol outperform the current practices in numerous ways. The proposed approach aims to achieve a balanced cluster distribution, extended longevity, and higher energy efficiency than standard procedures. Compared to existing methods, the suggested technique enhances network lifetime, throughput, latency, delivery of the packet, and energy optimization while maintaining satisfactory performance.
... 38,39 Its primary purpose is to utilize a database in which the data points are grouped into various sections to predict a new sample point. 40 It gives predictions just-in-time by measuring the similarity between each training instance and input sample. It is usually applied for its better calculation time and ease of interpretation. ...
In recommender systems, Collaborative Filtering (CF) plays an essential role in promoting recommendation services. The conventional CF approach has limitations, namely data sparsity and cold-start. The matrix decomposition approach is demonstrated to be one of the effective approaches used in developing recommendation systems. This paper presents a new approach that uses CF and Singular Value Decomposition (SVD)[Formula: see text] for implementing a recommendation system. Therefore, this work is an attempt to extend the existing recommendation systems by (i) finding similarity between user and item from rating matrices using cosine similarity; (ii) predicting missing ratings using a matrix decomposition approach, and (iii) recommending top-N user-preferred items. The recommender system’s performance is evaluated considering Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). Performance evaluation is accomplished by comparing the systems developed using CF in combination with six different algorithms, namely SVD, SVD[Formula: see text], Co-Clustering, KNNBasic, KNNBaseline, and KNNWithMeans. We have experimented using MovieLens 100[Formula: see text]K, MovieLens 1[Formula: see text]M, and BookCrossing datasets. The results prove that the proposed approach gives a lesser error rate when cross-validation ([Formula: see text]) is performed. The experimental results show that the lowest error rate is achieved with MovieLens 100[Formula: see text]K dataset ([Formula: see text], [Formula: see text]). The proposed approach also alleviates the sparsity and cold-start problems and recommends the relevant items.
Recommendation systems (RSs) belong to a category of information filtering systems designed to predict the “ranking” or “preference” that users will give to a particular item. RSs are automated instruments and strategies that assist and increase the decision-making process by aggregating the views of individuals and guiding them to suitable recipients. RSs are extensively utilized in various domains, including e-commerce, social networking sites, and entertainment, and impact everyone’s everyday life. The systems are designed to assist the user by proposing the items that are appropriate for him or her without requiring them to undergo the lengthy, time-consuming, and complex process of selecting from a wide selection of items that can number in the thousands or millions. The major aim of making recommendations based on the user’s interests is to minimize human work. Models and algorithms are expected to catch different user preferences and mostly identify non-dependencies between them and the multitude of items to provide personalization. In addition, this problem is compounded by real data criteria and ambitious real-time requirements. Many difficulties arise when developing and operating RSs. Therefore, it is compulsory to address them and design a system in which they become mitigated or tolerable. Sparsity, Cold Start, and Scalability are a few challenges when a user develops a recommendation system. The pervasive use of deep learning has demonstrated its power in solving complicated tasks more efficiently than conventional techniques. This paper seeks to stimulate advancements in RSs by providing a thorough summary of recent research on recommendation systems using deep learning. The surveyed articles are categorized using a taxonomy of recommendation systems that are offered. Based on the analysis of the evaluated works and the stated potential solutions, open problems are highlighted.
Clinicians and administrators are applying Artificial Intelligence (AI) Techniques widely as the promising results of their applications in the healthcare have been established. The meaningful impact of the AI applications will be limited unless it is coherently applied with human diagnosis and inputs from specialist clinician. This will help to address limitations and take advantage of the promises of the AI techniques. Machine Learning is one of the AI technique that finds high relevance in the medicine and health care. This review provides an overall glimpse of current practices and research outcomes of the application of the AI techniques in the healthcare and medical practices. It further describes Machine Learning Techniques in disease prediction and scope for food formulations for combatting disease.
The internet is the home for huge volumes of useful data that is constantly being created making it difficult for users to find information relevant to them. Recommendation System is a special type of information filtering system adapted by online vendors to provide recommendations to their customers based on their requirements. Collaborative filtering is one of the most widely used recommendation systems; unfortunately, it is prone to shilling/profile injection attacks. Such attacks alter the recommendation process to promote or demote a particular product. Over the years, multiple attack models and detection techniques have been developed to mitigate the problem. This paper aims to be a comprehensive survey of the shilling attack models, detection attributes, and detection algorithms. Additionally, we unravel and classify the intrinsic traits of the injected profiles that are exploited by the detection algorithms, which has not been explored in previous works. We also briefly discuss recent works in the development of robust algorithms that alleviate the impact of shilling attacks, attacks on multi-criteria systems, and intrinsic feedback based collaborative filtering methods.
Recommender System guides the users to choose objects from variety of possible options in personalized manner. Broadly, there are two categories of recommender systems i.e. content based and collaborative filtering based. These systems suggest the items based on the interest of the customers in the past. They personalize the information by using relevant information. These systems are used in various domains like recommending movies, products to purchase, restaurants, places to visit, etc. This chapter deliberates the concepts of content‐based recommender systems by including distinct features in their design and implementation. High level architecture and applications of these systems in various domains are also presented in this chapter.
In this paper, the use of TF-IDF stands for (term frequency-inverse document frequency) is discussed in examining the relevance of key-words to documents in corpus. The study is focused on how the algorithm can be applied on number of documents. First, the working principle and steps which should be followed for implementation of TF-IDF are elaborated. Secondly, in order to verify the findings from executing the algorithm, results are presented, then strengths and weaknesses of TD-IDF algorithm are compared. This paper also talked about how such weaknesses can be tackled. Finally, the work is summarized and the future research directions are discussed.
Humans are considered to be the most intelligent species on the mother earth and are inherently more health conscious. Since Centuries mankind has discovered various proven healthcare systems. To automate the process and predict diseases more accurately machine learn-ing methods are gaining popularity in research community. Machine Learning methods facilitate development of the intelligence into a machine, so that it can perform better in the future using the learned experience. Machine learning methods application on electronic health record dataset could provide valuable information and predication of health risks. The aim of this research review paper are four-fold: i) serve as a guideline for researchers who are new to machine learning area and want to contribute to it, ii) provide state-of-the-art survey of machine learning, iii) application of machine learning techniques in the health prediction, and iv) provides further research directions required into health prediction system using machine learning.
Machine learning is a pervasive development at the intersection of statistics and computer science. While it can benefit many data-related applications, the technical nature of the research literature and the corresponding algorithms slows down its adoption. Scikit-learn is an open-source software project that aims at making machine learning accessible to all, whether it be in academia or in industry. It benefits from the general-purpose Python language, which is both broadly adopted in the scientific world, and supported by a thriving ecosystem of contributors. Here we give a quick introduction to scikit-learn as well as to machine-learning basics.
In recent years, the role of cloud computing in the information technology (IT) industry has been growing steadily, partly because IT developers are becoming empowered to provision, manage, and adjust IT resources through simple application programming interfaces. This enables the easy creation and manipulation of software defined environments consisting of server, storage, network, software, and the supporting management functions. Model-based deployment and management technologies that are based on upcoming standards such as the Topology and Orchestration Specification for Cloud Applications (TOSCA) of the Organization for the Advancement of Structured Information Standards (OASIS) allow for describing models of different kinds of workloads such as topology patterns and management plans. The term “pattern” refers to a description of components of a service, e.g., an application or infrastructure, and their relationships. This approach provides a flexible combination of declarative and imperative models of workloads that can be seamlessly integrated with automation frameworks such as Chef for task-level automation as well as management services to drive resource orchestration and optimization. In this paper, we explain how the aforementioned concepts are used as foundational elements of the software defined environment. We use an actual example motivated by a real customer scenario, and we explain the relation to the IBM Cloud Computing Reference Architecture and how it is applied within IBM cloud implementations.
We describe an implementation of the well-known apriori algorithm for the induction of association rules [Agrawal et al. (1993), Agrawal et al. (1996)] that is based on the concept of a prefix tree. While the idea to use this type of data structure is not new, there are several ways to organize the nodes of such a tree, to encode the items, and to organize the transactions, which may be used in order to minimize the time needed to find the frequent itemsets as well as to reduce the amount of memory needed to store the counters. Consequently, our emphasis is less on concepts, but on implementation issues, which, however, can make a considerable difference in applications.
Data Science (Tukey in Exploratory data analysis: Past, present, and future. Princeton University Press, 1993; Naur in Communications of ACM 9:485, 1966; Hayashi et al. in Data Science, Classification and Related Methods, Proceedings of Fifth Conference of the International Federation of Classification Societies (IFCS—96). Kobe, Japan, 1998) is fast emerging as the inter-disciplinary field that converges in specialist professionals, domain expertise, data modelling expertise, statistical expertise and computer science. Data scientists transform raw data to intelligence through a systematic process of data understanding and model building based on their understanding of the data. However, to become effective in achieving their objective of transforming raw data to intelligence, they are challenged with a disjointed collection of tools, processes and unsatisfactory data acquisition and curation techniques.
The evidential K nearest neighbor classifier is based on discounting evidence from learning instances in a neighborhood of the pattern to be classified. To adapt the method to partially supervised data, we propose to replace the classical discounting operation by contextual discounting, a more complex operation based on as many discount rates as classes. The parameters of the method are tuned by maximizing the evidential likelihood, an extended notion of likelihood based on uncertain data. The resulting classifier is shown to outperform alternative methods in partially supervised learning tasks.
Development of technology in educational field brings the easier ways through the variety of facilitation for learning process, sharing files, giving assignment and assessment. Automated Essay Scoring (AES) is one of the development systems for determining a score automatically from text document source to facilitate the correction and scoring by utilizing applications that run on the computer. AES process is used to help the lecturers to score efficiently and effectively. Besides it can reduce the subjectivity scoring problem. However, implementation of AES depends on many factors and cases, such as language and mechanism of scoring process especially for essay scoring. A number of methods implemented for weighting the terms from document and reaching the solutions for handling comparative level between documents answer and expert's document still defined. In this research, we implemented the weighting of Term Frequency — Inverse Document Frequency (TF-IDF) method and Cosine Similarity with the measuring degree concept of similarity terms in a document. Tests carried out on a number of Indonesian text-based documents that have gone through the stage of pre-processing for data extraction purposes. This process results is in a ranking of the document weight that have closesness match level with expert's document.
Data mining is the process which is used to analyze the large database to find the useful pattern. Data mining can be used to learn about student's behavior from data collected using the course management system such as Moodle (Modular Object-Oriented Developmental Learning Environment). Here in this paper we show how data mining techniques such as clustering and association rule algorithm is useful in Course Recommendation System which recommends the course to the student based on choice of other students for particular set of courses collected from Moodle. As a result of Course Recommendation System, we can recommend to new student who has recently enrolled for some course e.g. Operating System, the new course to be opted e.g. Distributed System. Our approach uses combination of clustering technique - Simple K-means and association rule algorithm - Apriori and finds the result. These results were compared with the results of open source data mining tool-Weka. The result obtained using combined approach matches with real world interdependencies among the courses. Other combinations of clustering and association rule algorithms are also discussed here to select the best combination. This Course Recommendation System could help in building intelligent recommender system. This approach of recommending courses to new students can be immensely be useful in "MOOC (Massively Open Online Courses)".
A variety of prediction methods are used to relate high-dimensional genome data with a clinical outcome using a prediction model. Once a prediction model is developed from a data set, it should be validated using a resampling method or an independent data set. Although the existing prediction methods have been intensively evaluated by many investigators, there has not been a comprehensive study investigating the performance of the validation methods, especially with a survival clinical outcome. Understanding the properties of the various validation methods can allow researchers to perform more powerful validations while controlling for type I error. In addition, sample size calculation strategy based on these validation methods is lacking. We conduct extensive simulations to examine the statistical properties of these validation strategies. In both simulations and a real data example, we have found that 10-fold cross-validation with permutation gave the best power while controlling type I error close to the nominal level. Based on this, we have also developed a sample size calculation method that will be used to design a validation study with a user-chosen combination of prediction. Microarray and genome-wide association studies data are used as illustrations. The power calculation method in this presentation can be used for the design of any biomedical studies involving high-dimensional data and survival outcomes.
This chapter discusses content-based recommendation systems, i.e., systems that recommend an item to a user based upon a description of the item and a profile of the user's interests. Content-based recommendation systems may be used in a variety of domains ranging from recommending web pages, news articles, restaurants, television programs, and items for sale. Although the details of various systems differ, content-based recommendation systems share in common a means for describing the items that may be recommended, a means for creating a profile of the user that describes the types of items the user likes, and a means of comparing items to the user profile to determine what to recommend. The profile is often created and updated automatically in response to feedback on the desirability of items that have been presented to the user.
We developed a decision support system that helps doctors select appropriate first-line drugs. The system classifies patients' abilities to protect themselves from infectious diseases as a risk level for infection. In an evaluation of the prototype system, the risk level it determined correlated with the decisions of specialists. The system is very effective and convenient for doctors to use.
An evidential K-nearest neighbor classifier based on contextual discounting and likelihood maximization
- O Kanjanatarakul
- S Kuson
- T Denoeux
Kanjanatarakul, O., Kuson, S., Denoeux, T.: An evidential K-nearest neighbor classifier
based on contextual discounting and likelihood maximization. In: Destercke, S., Denoeux,
T., Cuzzolin, F., Martin, A. (eds.) BELIEF 2018. LNCS (LNAI), vol. 11069, pp. 155-162.